318 Intelligence and Miscellaneous Articles. 



so close together. This region also shows several triplets in which 

 one lateral component is apparently more intense than the other, 

 but in every such case the whole triplet is so very faint that I took 

 little notice of the phenomenon until my attention was called to 

 it by Dr. Zeeman's paper in the January number of the Astro- 

 physical Journal, in which he showed that such cases were due to 

 superpositions of other lines. 



In the experiments with magnesium, cadmium, and zinc, the 

 field used was in most cases about 25,000 c. G. s. units, and I 

 think fairly uniform, for the faces of the polepieces were nearly 

 flat and so close together that they left only a space about one-eighth 

 of an inch wide for the spark. 



The most interesting line in the spectrum of magnesium is 

 3832*45, one of a group of three very strong lines belonging to 

 the first subordinate series. This line has at least five components, 

 viz., a very strong central one (which may be further resolved in a 

 stronger field), and two lines symmetrically situated on either side of 

 it. Of these latter, the two extremes, which are very faint, are 

 like the central component due to vibrations normal to the field, 

 while the other pair are due to vibrations along the field. The 

 separation of the faint extremes is '60 unit, that of the inner 

 pair -23 unit. The other two lines of the group, 3829*5 and 

 3838-4, are triplets, the former sharp, the latter rather diffuse. 



In the Philosophical Magazine for February, Dr. T. Preston 

 gives a relationship between the separation for certain corre- 

 sponding lines of magnesium, cadmium, and zinc. He finds that the 



value of — is 18 for the lines Mg 5183-8, Cd 5086-0, Zn 4810-7: 



11-5 for Mg 5172-8, Cd 4800-1, and Zn 4722-3; and 10 for 



Mg 5167-5, Cd 4678-4, and Zn 4680-4. Moreover, the first three 



lines are diffuse triplets, the second three are quadruplets, and 



the third three are sharp triplets. My plates show very fair 



\ 2 

 agreement with Dr. Preston's results. My values of — are -39 

 ° AX 



for Mg 5183-8, -34 for Cd 5086-0, -35 for Zn 4810-7, '21 for 

 Mg 5172-8, -22 for Cd 4800-1 and Zn 4722-3, and about -19 for 

 Mg 5167-5, Cd 4678-4, and Zn 46S0-4 ; or approximately -36, -22,_ 

 and -19 for Preston's three groups, which are nearly proportional 

 to his figures. Only proportional agreement is to be expected, 

 since the field used was different in the two cases. His statements 

 in regard to the physical character of the lines are also confirmed 

 by my plates. 



In' the report of last spring in this Circular and in the Astro- 

 physical Journal, it was shown that the iron lines could be arranged 

 in groups with regard to the magnitude of the separation which 

 coincided with those found by Humphreys with regard to the shift 

 due to pressure, at least so far as the spectrum had been examined. 

 This, however, is certainly not the case with magnesium, cadmium, 

 and zinc. Por instance, the pressure-shift was found tp be 

 approximately -the same for the three 6-lines of magnesium, but 



